update readme

README.md

@@ -5,10 +5,11 @@ A neural net training interface based on TensorFlow.
 [![ReadTheDoc](https://readthedocs.org/projects/tensorpack/badge/?version=latest)](http://tensorpack.readthedocs.io/en/latest/index.html)
 [![Gitter chat](https://badges.gitter.im/gitterHQ/gitter.png)](https://gitter.im/tensorpack/users)
 
-See some [examples](examples) to learn about the framework:
+See some [examples](examples) to learn about the framework. Everything runs on multiple GPUs, because why not?
 
 ### Vision:
-+ [Multi-GPU training of ResNet on ImageNet](examples/ResNet)
++ [Train ResNet/SE-ResNet on ImageNet](examples/ResNet)
++ [Train Faster-RCNN on COCO object detection](examples/FasterRCNN)
 + [Generative Adversarial Network(GAN) variants](examples/GAN), including DCGAN, InfoGAN, Conditional GAN, WGAN, BEGAN, DiscoGAN, Image to Image, CycleGAN.
 + [DoReFa-Net: train binary / low-bitwidth CNN on ImageNet](examples/DoReFa-Net)
 + [Fully-convolutional Network for Holistically-Nested Edge Detection(HED)](examples/HED)

examples/FasterRCNN/train.py

@@ -252,6 +252,7 @@ if __name__ == '__main__':
     if args.evaluate is not None:
         assert args.evaluate.endswith('.json')
         assert args.load
+        os.environ['TF_CUDNN_USE_AUTOTUNE'] = '0'
         offline_evaluate(args.load, args.evaluate)
         sys.exit()
     if args.predict is not None:

examples/README.md

@@ -10,17 +10,18 @@ Training examples with __reproducible__ and meaningful performance.
 
 ## Vision:
 + [A tiny SVHN ConvNet with 97.8% accuracy](svhn-digit-convnet.py)
-+ Multi-GPU training of [ResNet](ResNet) and [ShuffleNet](ShuffleNet) on ImageNet
++ Train [ResNet](ResNet) and [ShuffleNet](ShuffleNet) on ImageNet
++ [Train ResNet50-Faster-RCNN on COCO](FasterRCNN)
 + [DoReFa-Net: training binary / low-bitwidth CNN on ImageNet](DoReFa-Net)
-+ [Generative Adversarial Network(GAN) variants](GAN), including DCGAN, InfoGAN, Conditional GAN, WGAN, BEGAN, DiscoGAN, Image to Image, CycleGAN.
++ [Generative Adversarial Network(GAN) variants](GAN), including DCGAN, InfoGAN, Conditional GAN, WGAN, BEGAN, DiscoGAN, Image to Image, CycleGAN
 + [Inception-BN and InceptionV3](Inception)
 + [Fully-convolutional Network for Holistically-Nested Edge Detection(HED)](HED)
 + [Spatial Transformer Networks on MNIST addition](SpatialTransformer)
 + [Visualize CNN saliency maps](Saliency)
 + [Similarity learning on MNIST](SimilarityLearning)
 + Learn steering filters with [Dynamic Filter Networks](DynamicFilterNetwork)
-+ Load a pre-trained [AlexNet](load-alexnet.py) or [VGG16](load-vgg16.py) model.
-+ Load a pre-trained [Convolutional Pose Machines](ConvolutionalPoseMachines/).
++ Load a pre-trained [AlexNet](load-alexnet.py) or [VGG16](load-vgg16.py) model
++ Load a pre-trained [Convolutional Pose Machines](ConvolutionalPoseMachines/)
 
 ## Reinforcement Learning:
 + [Deep Q-Network(DQN) variants on Atari games](DeepQNetwork), including DQN, DoubleDQN, DuelingDQN.